ZERITH 2.5 model weight attribution Model: ZERITH 2.5 LIBERO Publisher and attribution: Zerith Robotics Copyright 2026 Zerith Robotics, for its own contributions. Original model repository: https://huggingface.co/OpenZerith/ZERITH-2.5-LIBERO Original weight revision: ad1659660704207d64c8b05f9a6b6f75eca224b5 Model SHA256: bdb9cb042ff87f437c5e1b184a710e5c127d89afdc4be28756483311d3e3970a Zerith-owned rights in these model weights are licensed under Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0); see LICENSE. No commercial-use permission is granted. The license text is unchanged. This metadata update adds licensing/attribution and does not change tensors. Upstream base model: Wan-AI/Wan2.2-TI2V-5B by the Wan Team Model repository: https://huggingface.co/Wan-AI/Wan2.2-TI2V-5B Model revision: 921dbaf3f1674a56f47e83fb80a34bac8a8f203e Source repository: https://github.com/Wan-Video/Wan2.2 Source revision: 42bf4cfaa384bc21833865abc2f9e6c0e67233dc Upstream license: Apache License 2.0; see LICENSES/Apache-2.0.txt. The pretrained base parameters were adapted during ZERITH 2.5 training. Original upstream rights remain under their own terms; this notice does not relicense those rights or restrict independent use of upstream material. External Wan VAE, UMT5 and VGGT encoder assets are obtained separately and are not bundled in this policy weight release. Preserve their respective terms and attribution when obtaining or redistributing them. See dependencies.json.